Abstract

Double walled carbon nanotube (DWCNT)/N,Pd codoped TiO2 nanocomposites were prepared by a modified sol-gel method and characterised using FTIR, Raman spectroscopy, TGA, DRUV-Vis, XRD, SEM, and TEM analyses. TEM images showed unique pearl-bead-necklace structured morphologies at higher DWCNT ratios. The nanocomposite materials showed characteristic anatase TiO2 Raman bands in addition to the carbon nanotube D and G bands. Red shifts in the UV-Vis absorption edge were observed at low DWCNT percentages. The photocatalytic activity of DWCNT/N,Pd TiO2 nanocomposite was evaluated by the photocatalytic degradation of eosin yellow under simulated solar light irradiation and the 2% DWCNT/N,Pd TiO2 nanocomposite showed the highest photoactivity while the 20% DWCNT/N,Pd TiO2 hybrid was the least efficient. The photocatalytic enhancement was attributed to the synergistic effects of the supporting and electron channeling role of the DWCNTs as well as the electron trapping effects of the platinum group metal. These phenomena favour the separation of the photogenerated electron-hole pairs, reducing their recombination rate, which consequently lead to significantly enhanced photoactivity.

1. Introduction

The discovery of carbon nanotubes (CNTs) in the early 1990s revolutionalised research in a myriad of R&D fields such as nanoelectronics, biosensors, intercalation materials, supercapacitors, drug delivery, polymer composites, and catalysis [1]. The unique and outstanding mechanical and electronic properties of CNTs have provided new impetus in their potential application in semiconductor nanocomposites as photocatalysts for degradation of organic pollutant and bacteria as well as water splitting to generate hydrogen gas [24]. Their one-dimensional structure allows electrons to be conducted without being scattered, a phenomenon known as ballistic transport which permits the nanotube to conduct without dissipating energy as heat [5]. This property is quite enticing in the fabrication of CNT-semiconductor hybrids in which electron channeling through the CNT is desirable. The semiconductor characteristics of CNTs may also induce changes in the reactivity of supported semiconductor nanoparticles by electron transfer. Reports on Coulombic charge transfer between Pt and Au metal nanoparticles supported on the SWCNT and the semiconducting SWCNTs have been made [6]. The surface area of CNTs may range from over 1000 m2/g for single walled carbon nanotubes (SWCNTs) to 700–800 m2/g for double walled carbon nanotubes (DWCNTs) and 200–300 m2/g for multiwall carbon nanotubes (MWCNTs). The large surface area provides spatial confinement of any attached nanoparticles and large supporting surface areas, leading to faster rates of redox reactions. SWCNTs and DWCNTs are thus preferable candidates for a number of applications; however, the lack of low cost mass production is still a limit to their practical use [7].

Titanium dioxide (TiO2) has emerged as the most preferable among many photocatalysts because of its many desirable properties such as nontoxicity, strong oxidizing power, low cost, ready availability, and long term photostability [8]. However, application of TiO2 is hindered by its large band gap (3.2 eV), which translates to use of high energy UV radiation for activation. Various innovations aimed at improving the photoefficiency of TiO2 have been attempted, including modifications with metals [9, 10] and nonmetals [11, 12]. Simultaneous introduction of nonmetals and metals as TiO2 dopants (codoping) has also been attempted in an effort to improve visible light responsiveness and reduction in electron-hole recombination of the resulting codoped TiO2 [1315].

CNTs/TiO2 nanocomposites have been reported to promote the separation of the electron-hole charges generated upon irradiation as well as shift the absorption edge of the nanocomposite to the visible region [16]. However, most studies have shown nonuniform coatings of TiO2 on the CNTs, with some bare CNTs surfaces and random aggregates of TiO2 on the CNT surface, thus compromising the efficiency of the nanocomposite [17]. Recent studies on CNT/TiO2 nanocomposites are focused mainly on MWCNT/TiO2 and SWCNT/TiO2 hybrids [1822]. In this work, a simple sol-gel technique is used to synthesize DWCNT/N,Pd codoped TiO2 nanocomposites by anchoring N,Pd codoped TiO2 on the surfaces of DWCNTs. The photocatalytic activity of the synthesized nanocomposite was evaluated by photocatalytic degradation of eosin yellow under simulated solar light irradiation.

2. Experimental

2.1. Synthesis, Purification, and Functionalisation of DWCNTs

A catalyst with the elemental composition Mg0.94Co0.05Mo0.01O was prepared through combustion of Mg and Co nitrates and hexaammonium heptamolybdate tetrahydrate in the presence of citric acid [23]. In a typical experiment, Mg(NO3)26H2O (70.53 g), Co(NO3)26H2O (4.29 g), (NH4)6Mo7O244H2O (3.72 g), and citric acid (170.16 g) were dissolved in deionised water and the solution stirred at 150°C until the solution turned purplish. The solution was then placed in a crucible and heated in a furnace at 550°C overnight. The resulting fluffy khaki brown oxide was finely ground and used as a catalyst in the synthesis of DWCNTs. The catalyst (0.5 g) was loaded into a ceramic boat which was then placed in a quartz tube connected to the gases and heated in an electric furnace at 1000°C for 2 h. Hydrogen gas (H2) and methane gas (CH4) were passed through the catalyst at constant flow rates of 50 mL/min and 200 mL/min, respectively, for 2 h and then the furnace and methane gas were switched off and hydrogen flow was maintained until the furnace cooled down to about 400°C. The pristine DWCNTs containing carbonaceous impurities and unreacted metal catalyst particles were sonicated in concentrated HCl (150 mL) for 3 h and washed several times with deionised water, filtered, and then dried in an oven at 90°C. The purified DWCNTs (2 g) were functionalised by sonication in a mixture of concentrated H2SO4 and HNO3 (1 : 3 by volume) at room temperature for 3 h. The oxidised DWCNTs were then washed several times with deionised water, filtered until the filtrate pH was 7, and then dried overnight at 90°C.

2.2. Preparation of DWCNT/N,Pd Codoped TiO2 Nanocomposites

Functionalised DWCNTs were dispersed in a solution of 2-propanol C3H8O (50 mL) (99.8%, Sigma Aldrich, Germany) and titanium isopropoxide (TTIP) (10 mL) (97%, Sigma Aldrich, Germany). The mixture was sonicated for 30 minutes to improve CNT dispersion. Palladium diamine dichloride, Pd(NH3)2Cl2, (45% Pd, PGM Chemicals, RSA) was then dissolved in 12 mL of aqueous ammonia (25%, Merck, Germany) to give a Pd : Ti ratio of 0.5% and added dropwise to the DWCNT/2-propanol/titanium isopropoxide mixture with vigorous stirring for 30 minutes and the precipitate was left to age for 4 hours. The suspension was then oven dried overnight at 80°C. The resulting powder was calcined in a furnace for 2 h at 450°C before characterisation. Nanocomposites containing 0.5%, 1%, 2%, 5%, 10%, and 20% DWCNTs were prepared by weighing appropriate amounts of oxidised DWCNTs and denoted NC-1 to NC-6, respectively.

2.3. Characterisation of the Materials

FTIR analysis was performed to probe the functional groups present on the materials. A Perkin Elmer Spectrum 100 was used. The nature of the DWCNTs and the nanocomposites was confirmed using a Raman microscope (Perkin Elmer RamanMicro 200). X-ray diffraction (XRD) analysis was performed at room temperature on a Philips PANanalytical X’pert operated at 40 kV and 40 mA. Optical properties were investigated in the reflectance mode on a Shimadzu UV-2540. Morphological properties were observed on a transmission electron microscope (TEM) (Tecnai G2 Spirit) equipped with an INCA EDS analyser and scanning electron microscope (SEM) (NOVA FEI/FIB). Thermal properties were evaluated by thermogravimetric analysis (TGA) (Perkin Elmer TGA 4000) over a temperature range of 30°C–800°C.

2.4. Photocatalytic Activity Tests

A portion of the NC (0.1 g) was suspended in the dye (eosin yellow) solution (100 mL and 100 ppm) and stirred in the dark to allow for adsorption equilibrium for about an hour. An Oriel Newport 500 W Xenon lamp solar simulator equipped with an Air Mass 1.5 Global Spectral Filter was employed for sample irradiation. The power output was measured with an Oriel PV reference cell. Aliquots (2 mL) of the suspension were sampled at 15-minute intervals and filtered through a syringe filter and the dye concentration was determined using UV-Vis spectroscopy (Shimadzu UV-2450) at λ = 515 nm.

3. Results and Discussion

FTIR analysis confirmed the presence of OH groups centred at 3351 cm−1 due to carboxyl functionalisation of the DWCNTs. The band at 1461 cm−1 present in both pristine and oxidised DWCNTs was ascribed to the C=C stretching vibrations. The two bands at 1120 cm−1 and 1594 cm−1 correspond to C–O and C=O stretching of carboxyl groups, further confirming the DWCNTs were –COOH functionalised (Figure 1(a)). These become anchor sites for attachment of TiO2 nanoparticles forming C–O–Ti linkages and thus the NC spectra were largely similar with the main peaks below 550 cm−1, ascribed to the Ti–O–Ti and Ti–O–C vibrations (Figure 1(b)). Broad band peaks centred at around 3390 cm−1 confirmed the OH stretching of physisorbed water on the nanocomposite surfaces. Sharp bands at around 1630 cm−1 were attributed to the OH bending modes of the water molecules. However, since ammonium hydroxide was used as the N precursor for the N,Pd codoped TiO2, it also forms linkages with the available functional groups on the DWCNTs forming the N=C=O linkages with characteristic vibrations at 2350 cm−1. Peaks associated with the functionalisation of DWCNTs were not observed in the nanocomposites. This could indicate formation of linkages between the DWCNTs and the TiO2 which occurs through –COOH functionalities.

Raman spectra of pristine and oxidised DWCNTs showed radial breathing mode (RBM) vibrations in the low frequency region (100–400 cm−1) in addition to the D (1305 cm−1) and G bands (1582 cm−1) (Figure 2(a)) [24]. The nanocomposites showed characteristic peaks for DWCNTs as well as anatase TiO2 (Figure 2(b)). The D and G bands centred at 1299 cm−1 and 1572 cm−1, respectively, were observed for NC-4 to NC-6. The slight decrease in the nanocomposite peak frequencies compared to the oxidised DWCNTs could indicate formation of Ti–O bonds with oxygen functional groups on the chemically oxidised DWCNTs. The lower frequency Raman region is dominated by the TiO2 bands. Peaks centred at 151 cm−1, 395 cm−1, 510 cm−1, and 636 cm−1 can be ascribed to the , , unresolved , and Raman allowed modes of anatase TiO2, respectively [25, 26].

X-ray diffraction patterns at 2θ values of 25.29, 37.83, 48.07, 54.14, 55.07, 62.64, 68.94, 70.14, and 75.11 can be indexed to the (101), (004), (200), (105), (211), (204), (116), (220), and (215) crystal planes of the anatase phase of TiO2, respectively [27, 28] (Figure 3). The DWCNT peaks (expected at 2θ = 25.6°) could not be detected due to possible overlap with the high intensity TiO2    peak [26].

Crystallite sizes () were estimated from the Scherrer equation: where is the X-ray wavelength, is the broadening of the peak, is the angle of diffraction, and is a constant [15, 27]. The anatase (101) peak at = 25.3° was used to estimate the crystallite sizes. There were significant differences in crystallite sizes with increasing amounts of DWCNTs in the nanocomposites (Table 1). Similar studies on Fe/TiO2-MWCNT composites reported average crystallite size of 5 nm [29].

Optical properties of the nanocomposites are shown in Figure 4(a). A fundamental absorption edge in the UV region, attributed to charge transfer from the valence to the conduction band of the TiO2, is observed for all samples [30]. The strong UV absorption is accompanied by tailing well into the visible region which decreases in the midrange with increasing amount of DWCNTs in the nanocomposites. The strongest absorption in the visible region was observed for NC-1 centred at around 500 nm. Increasing the amount of DWCNTs in the nanocomposites resulted in a decrease in the absorption in the region around 500 nm but an increase was observed between 700 nm and 900 nm. Increase in visible light absorption can be explained by N doping which forms interband gap states in TiO2, resulting in reduced band gap and consequently shifting the absorption edge to the visible region. The presence of Pd is also associated with absorption in the visible region due to d-d electron transitions [15]. CNTs are known to absorb well in the entire visible region, so their presence in the nanocomposites also contributes to enhanced visible light absorption. Controlling the composition ratio of DWCNTs in these nanocomposites is crucial for optimal synergistic effect between the DWCNTs and the codoped TiO2. The band gap () of the materials was estimated from the Tauc plots by plotting []n against (Figure 4(b)). An increase in band gap energy with increasing percentages of DWCNTs in the nanocomposites (Table 1) was observed. Higher amounts of DWCNTs limit the available N and Pd for doping with TiO2 as some of these elements may become associated with the DWCNTs, causing gradual increase in band gap. Lower band gaps (increased visible light absorption) were only achieved at lower DWCNT ratios.

SEM and TEM images of the oxidised and TiO2 covered DWCNTs are shown in Figure 5. The DWCNTs appeared as long and narrow diameter tubes (Figures 5(a) and 5(b)). SEM images of the NC-5 show clusters of codoped TiO2 surrounding the DWCNTs to form rough tubular structures (Figure 5(c)) The diameter of the nanocomposites was visibly larger than that of the oxidised DWCNTs due to TiO2 coverage. The length of the nanocomposites was shorter than that of the oxidised DWCNTs suggesting possible shortening during the sol-gel processing. The formation of TiO2 aggregates confirmed the role of the DWCNTs as support for the deposition and growth or spatial confinement of the N,Pd codoped TiO2 clusters. TEM image of the nanocomposite (NC-5) shows TiO2 particles anchored onto the small diameter DWCNTs to form long pearl-bead-necklace-like structures (Figure 5(d)). The bead-like TiO2 particle coverage indicates strong contact with the DWCNTs. The average size of the anchored TiO2 nanoparticles is about 10 nm.

EDS analysis was performed on one of the nanocomposites (NC-3) to probe the elemental composition of the material (Figure 6). The presence of Ti, O, Pd, and C confirmed the existence of Pd doped TiO2 on the DWCNTs. The intensity of the C peak was much lower compared to the Ti and O peak intensities, suggesting near complete coverage of the DWCNTs by the TiO2 concurring with SEM and TEM observations. Pd showed the lowest peak intensity as expected because of the low levels (0.5%) used in preparing the materials. The Si could have originated as a contaminant from the glassware while the Cu signal came from the copper grid used as the sample support during TEM analysis.

Thermograms of the nanocomposites show a small weight loss (about 2%) below 100°C, attributed to the adsorbed water molecules on the materials (Figure 7). A weight loss at around 550°C can be attributed to the loss of the graphitic skeleton of DWCNTs in the nanocomposites. The total carbon loss at this temperature reveals the percentage DWCNTs in the respective nanocomposites and the residual percentage weight gives an approximate amount of N,Pd TiO2 in the nanocomposites. The percentage weight loss values at 550°C are lower than those predicted from the starting precursor materials. This may be due to the bundling together and aggregation of the DWCNTs, a prominent phenomenon at higher CNT percentages.

Dye degradation was evaluated from the measured concentrations of the sampled dye aliquots at 15-minute intervals using the following equation:where and are the concentrations of dye at reaction time and 0, respectively.

Reaction kinetics data were computed from the simplified Langmuir-Hinshelwood kinetics model, assuming first order kinetics:where is the apparent rate constant and is time.

Photocatalytic evaluation of the nanocomposites showed the highest activity for the NC-3 (Figure 8(a)). Initially there was an increase in degradation efficiency with increasing amount of DWCNT and then a decrease at much higher DWCNT percentages. Photocatalytic enhancement can be explained in terms of DWCNT/N,Pd TiO2 heterojunctions that help reduce the rate of the recombination of photogenerated charge carriers. There is also possibility of shifting of the TiO2 apparent Fermi level thereby allowing the utilisation of low energy photons for photo activation. Presence of DWCNTs promotes excellent charge transfer ability, high surface area for increased dye adsorption, and the formation of Ti–C bonds between the CNTs and TiO2, resulting in enhanced visible light absorption [3032].

DWCNTs themselves can also absorb photons, get excited, and inject electrons into the TiO2 conduction band, triggering the formation of highly reactive superoxide and hydroxyl radicals [33]. The decrease in photoactivity at higher DWCNT percentages correlated well with the increase in band gaps estimated from the Tauc plots. From the observed trend, it can be postulated that there is an optimum DWCNT/TiO2 ratio that offers the best conditions for the synergy between DWCNTs and N,Pd TiO2 in the nanocomposites. The photodegradation kinetics followed pseudo first order kinetics postulated from the Langmuir-Hinshelwood model (Figure 8(b)). Dye adsorption, degradation, and apparent rate constants are summarised in Table 2. In a study on CNT/TiO2 nanocomposites with different CNT loadings, the optimum CNT/TiO2 ratio was found to be in the range from 1.5% to 5% by weight [33]. A study on the degradation of acetone using TiO2/CNT nanocomposites gave an optimal amount of TiO2 and CNTs in the range of 10 : 1 and 5 : 1 [34].

A rate constant of 3.220 × 10−2 min−1 observed for NC-3 (2.0% DWCNT/N,Pd TiO2) is higher compared to the one reported for similar work performed using MWCNT/N,Pd TiO2 nanocomposites [15, 35]. However, both results reveal an improvement compared to commercial TiO2 (P25) which was used as a reference under similar conditions (Table 3). N doped and N,Pd codoped TiO2 performed better than the P25 with N,Pd codoped TiO2 showing even better performance than when incorporated into CNTs. The NC-3 showed a rather mixed order reaction kinetics shown by the S shaped curve. This can be attributed to the presence of both pearl-bead-necklace structures as well as formation of aggregates on the DWCNTs surface.

UV-Vis scans of the dye photodegradation ware plotted as a function of irradiation time for the NC-6 (Figure 9) and no new absorption peaks were observed. This confirmed that the dye was completely mineralised.

A mechanism for the photocatalytic enhancement was proposed based on substitution of O by N in the TiO2 lattice resulting in formation of interband gap states and therefore facilitating the use of less energetic visible light to activate the TiO2 (Figure 10). Pd becomes an electron sink due to the formation of a Schottky barrier at the semiconductor metal interface [2, 36]. DWCNTs act as effective electron transfer channels due to their high electrical conductivity. They exhibit high electron capture ability and high electron storage capacity and can act as super capacitors [37]. DWCNTs can also act as photosensitisers, because of their semiconductive properties, absorbing visible light radiation and transferring the photogenerated electrons into the conduction band of the N,Pd codoped TiO2 particles [38]. The resulting positively charged DWCNTs can capture electrons from the valence band of N,Pd codoped TiO2, thereby extending the lifetime of holes.

4. Conclusion

Anchoring N,Pd codoped TiO2 on DWCNTs led to changes in the morphology and optical properties of the composite materials resulting in enhanced photocatalytic activity of the nanocomposites towards degradation of eosin yellow. The ability of DWCNTs and Pd to act as electron sinks highlights their mediating roles in the charge transfer process. DWCNTs also reduce N,Pd codoped TiO2 agglomeration and induce synergistic effects on the photocatalytic performance. Photocatalytic enhancement was only observed at low DWCNT : Ti ratio of 2.0% by weight. DWCNT/N,Pd codoped TiO2 are promising candidates for addressing some of the fundamental issues in many practical environmental applications including removal of organic pollutants from water.

Competing Interests

The authors declare no conflict of interests regarding the publication of this article.

Acknowledgments

The authors acknowledge funding from the Nanotechnology and Water Sustainability Research Unit (NanoWS), University of South Africa (UNISA).